A critical issue for disaster response is group or collaborative computing in mobile environments. This application describes research to address three critical inter-related problems that inhibit the successful use of information technologies at disaster sites to manage medical care. Problem #1: Disaster sites often have a noisy and chaotic electromagnetic environment that makes wireless networks unreliable. Practical systems for disaster response must have architectures that preserve first responder work flows, including mobile group work, even during disruptions that might prevent connects to any central server. Problem #2: First responders typically arrive at the scene of disaster sites one company at a time and organize themselves based on the Incident Command System. Practical systems for disaster response must be interoperable and self-scaling with progressive growth of capabilities as first responders arrive at the scene. Problem #3: The capabilities of computer systems in disaster environments will change from moment to moment based on connectivity and infrastructure. How can we convey this information to users and design applications that will use this information in a seamless way to enhance work flow? This application describes research using the Wireless Internet Information System for Medical Response in Disasters (WIISARD) test bed to study how best to address these issues by creating self-scaling systems for disaster management. One aim will be to research models for transactions for systems that relax the requirement for data transactions to close simultaneously through the system and identify types of transaction models best suited to the domain of disaster response. Other work will examine how publish/subscribe protocols can be adapted to yield self-scaling systems. User oriented design and other design exercises will help refine the features of interfaces of core medical applications for disaster care that use these alternative transaction models and also expand and contract in scope of operations. Measurements of the effectiveness of models and innovations will be performed using the entire WIISARD test bed in large disaster drills with the San Diego Regional Metropolitan Medical Strike Team and Disaster Medical Assistant Team CA-4. Measures will include network and application performance, speed of transmission of information, and effects on clinical work flow. If successful, this research will overcome critical technical obstacles that prevent the effective use of computer systems to support medical care at disaster sites.